Dec 29 2024
ChatGPT Search Manipulated With Hidden Instructions via @sejournal, @martinibuster

New report claims that ChatGPT Search can be manipulated with hidden text featuring instructions telling ChatGPT Search how to respond to an answer Tests also showed that ChatGPT could be manipulated without the instructions, with just the hidden text.

ChatGPT Search Can Be Manipulated With Hidden Text

A report from The Guardian outlines how they used hidden text on a fake website to trick ChatGPT Search to show them a response from hidden text on the web page. Text is hidden when the font matches the background color of a page, like a white font on a white background.

They then asked ChatGPT Search to visit the website and answer a question based on the text on the site. ChatGPT Search browsed the site, indexed the hidden content and used it in the answer.

They first assessed ChatGPT using a non-exploit control page on a fake review website to test ChatGPT’s response. It read the reviews and returned a normal response.

Researchers at The Guardian next sent ChatGPT Search to a fake website that had instructions to give a positive review and ChatGPT Search followed the instructions and returned positive reviews.

The researchers did a third test with positive reviews written in hidden text but without instructions and ChatGPT Search again returned positive reviews.

This is how The Guardian explained it:

“…when hidden text included instructions to ChatGPT to return a favourable review, the response was always entirely positive. This was the case even when the page had negative reviews on it – the hidden text could be used to override the actual review score.

The simple inclusion of hidden text by third parties without instructions can also be used to ensure a positive assessment, with one test including extremely positive fake reviews which influenced the summary returned by ChatGPT.”

The above test is similar to a test of ChatGPT that computer science university professor did in March 2023 where he tricked ChatGPT to say that he was a time travel expert.

What these tests prove is that ChatGPT’s training data and the ChatGPT Search Bot ingest hidden text but can also be manipulated into following directions. The Guardian quotes a security expert saying that OpenAI was made aware of the exploit and that it might be fixed by the time the article is published.

Why Can AI Search Engines Be Manipulated?

One loophole in AI Search is a technology called RAG (Retrieval Augmented Generation), a technique that can fetch information from a search engine so that an AI can then use it for generating answers to questions from up to date and (presumably) authoritative sources. How do AI Search Engines determine authoritative web pages? Perplexity AI, for example, uses a modified version of PageRank in order to identify trustworthy web pages to cite in their AI search engine.

ChatGPT Search is based on Bing but it also has its own crawler that can fetch real-time information. It’s probably not unreasonable to speculate that if a site is included in Bing’s search index then it’s probably included within ChatGPT Search, which should protect ChatGPT Search from being influenced by hidden text. Presumably, sites with hidden text would be excluded from Bing’s search index. That said, it may be possible to cloak a website so that it shows different content to the ChatGPT Search Bot (an up to date list of OpenAI Search Crawler bots is available here).

Other Ways To Manipulate AI Search Engines

There are said to be other ways that researchers discovered last year that might still be effective (Read: Researchers Discover How To SEO For AI Search). In this research paper from last year the researchers tested nine strategies for influencing AI search engines:

Nine Strategies For Manipulating AI Search Engines

  1. Authoritative: Changing the writing style to be more persuasive in authoritative claims
  2. Keyword optimization: Adding more keywords from the search query
  3. Statistics Addition: Changing existing content to include statistics instead of interpretative information.
  4. Cite Sources (quoting reliable sources)
  5. Quotation Addition: Adding quotes and citation from high quality sources
  6. Easy-to-Understand: Making the content simpler to understand
  7. Fluency Optimization is about making the content more articulate
  8. Unique Words: Adding words that are less widely used, rare and unique but without changing the meaning of the content
  9. Technical Terms: This strategy adds both unique and technical terms wherever it makes sense to do so and without changing the meaning of the content

The researchers discovered that the first three strategies worked the best. Notably, adding keywords into web pages helped a lot.

ChatGPT Search Can Be Manipulated?

I overheard claims made at a recent search conference that Google AI Overviews could be manipulated to show certain big brand products in response to search queries. I didn’t verify whether that was true but the claim was made by a reliable and authoritative source. With regard to ChatGPT Search, I’ve noticed some interesting things about what sites it chooses to surface information and under what circumstances, which could be a way to influence rankings. So it’s not surprising that there are ranking loopholes in ChatGPT Search. AI Search is looking a lot of like the early days of traditional search.

Featured Image by Shutterstock/Antonello Marangi

Ecommerce MGMT 0 Comments
Dec 26 2024
Major Outage Hits OpenAI ChatGPT via @sejournal, @martinibuster

ChatGPT Search is down for users worldwide, becoming unresponsive at approximately 11 AM EST. Down detectors and reports on social media reflect that many users cannot access the site. OpenAI is reporting a major outage.

Major Outage At OpenAI

OpenAI is reporting a major outage is hitting it’s ChatGPT, API, and SORA services. The issue is described by OpenAI as caused by an upstream provider. Upstream provider is a vague description that can be compute-related on Microsoft’s side or a network related service.

Screenshot Of  Official ChatGPT Outage Status

OpenAI Sora Is Only Partially Affected

The major outage has hit both ChatGPT and the API but OpenAI’s Sora service is only partially affected.

Screenshot Shows Severity Of Outage

The major outage has hit both ChatGPT and the API but OpenAI’s Sora service is only partially affected.

The outage began by at least 11 AM and the problem was identified 18 minutes later. OpenAI continues to work on a solution an hour later.

This is a developing event and this article will be updated as more information becomes available.

Featured Image by Shutterstock/Tithi Luadthong

Ecommerce MGMT 0 Comments
Dec 22 2024
Study: ChatGPT & AI Tools Gain Ground In Search Market via @sejournal, @MattGSouthern

A new study by Previsible reveals significant changes in search behavior, with AI language models (LLMs) gaining traction as referral traffic sources.

The analysis of over 30 websites shows Perplexity and ChatGPT emerging as alternatives to traditional search engines.

David Bell, co-founder of Previsible, believes this indicates Google’s growth is at a standstill:

“Google is basically plateaued and has begun to have its search dominance degraded. The reason being is that people are starting to use ChatGPT, Claude, Co-pilot, Bing, all these different experiences to better solve their search intent.”

Here are some key points from the study. While it doesn’t provide a complete picture, it offers the best information available right now.

Key Findings

Referral Traffic

The study found that Perplexity and ChatGPT command approximately 37% of LLM referral traffic, while CoPilot and Gemini follow with 12-14% each.

Notably, the finance sector dominates LLM-driven traffic, accounting for 84% of all referrals analyzed.

In a video walkthrough of the study, Bell explains:

“Finance, in particular, has an outsized increase in traffic from language models. This could be due to Perplexity and other language models having integrations or relationships with different platforms that allow more direct access to users.”

Content Distribution

The study reveals that blog posts receive 77.35% of LLM referral traffic, followed by:

  • Homepage visits (9.04%)
  • News content (8.23%)
  • Guides (2.35%)

“Informational content still matters in the age of AI search,” Bell noted. However, he advises focusing on conversion rate optimization (CRO) and user journey, as product pages do not surface prominently in language models.

According to the study, product pages receive less than 0.5% of LLM referral traffic, suggesting challenges for ecommerce strategies.

Looking Ahead

LLM referral traffic currently represents 0.25% of total traffic for the most impacted sectors, though the study notes significant growth rates.

In the last 90 days of the study, Previsible found:

  • 900% growth in ChatGPT referrals for the events industry
  • 400%+ growth in ChatGPT traffic for e-commerce and finance sectors
  • Consistent growth across all models except CoPilot

Bell explains what this could mean for websites:

“If you extrapolate out, if you average all of these out, let’s say roughly 200% growth in organic or AI traffic every 90 days for the next 12 months, it can be up to 20% of overall traffic to a website.”

🚨 Free Tool Alert 🚨

Previsible has created a free Looker Studio dashboard to help businesses track website traffic from Language Learning Models (LLMs) like ChatGPT, Perplexity, Co-pilot, Gemini, and Claude.

You can select your Google Analytics 4 (GA4) account and a date range to view data.

The dashboard shows:

  • Organic Sessions: Total organic sessions during the selected time.
  • Total LLM Sessions: Number of sessions from LLMs, with percentage breakdowns.
  • LLM Traffic Over Time: A line graph showing LLM traffic trends, with separate lines for each LLM.
  • LLM Traffic by Landing Page: A table of top landing pages from LLM traffic, including sessions, percentages, average time on page, and comparison to site averages.

How This Helps

By analyzing this data, you can:

  • Assess the impact of LLMs on your traffic compared to organic traffic.
  • Identify which LLMs drive the most traffic and adjust your content.
  • Track LLM traffic growth over time and adapt your strategies.
  • Discover popular landing pages among LLM users and improve them for engagement.
  • Compare time spent by LLM users on each page to the site average to identify areas for improvement.

In Summary

Here are three key takeaways from the study:

  • Finance websites are seeing the strongest LLM referral activity, with blog content receiving the majority of visits
  • Product pages rarely surface in LLM results, suggesting the need for adjusted e-commerce strategies
  • Growth rates are significant, potentially reaching 20% of total traffic within a year if current trends continue

When looking at these trends, it’s important to keep a balanced approach to getting traffic and optimizing your strategies.

Don’t pursue AI traffic if it could hurt your sales.

AI language models are becoming new sources of website traffic. However, they currently make up only about 0.25% of overall traffic in the sectors that are most affected.

It will be interesting to see how this number changes by next year.

Featured Image: Koshiro K/Shutterstock

Ecommerce MGMT 0 Comments
Dec 19 2024
Cut The Malarkey. Speaking Frankly About AI Search & SEO via @sejournal, @martinibuster

Search marketing is undergoing dramatic changes, with many debating whether SEO is on its way out as AI Search rises in popularity. What follows is a candid assessment of what is going on with SEO and search engines today.

An SEO School Shuts Down

An SEO school by a group called Authority Hackers recently announced their closure, emphasizing that it’s not because SEO is dead but due to the collapse of the content site model. They cited three reasons for this situation. The following is not about the SEO school, that’s just a symptom of something important going on today.

1. Google Updates is one of the reasons cited for the decline of the content site model. Here’s the candid part: If the Google updates killed your publishing site, that’s kind of the red flag that there’s something about the SEO that needs examination.

Here’s the frank part: Google’s updates have generally crushed websites that begin with keyword research, are followed by stealing content ideas from competitors and scraping Google’s SERPs for more keyword phrases. That’s not audience research, that’s search engine research. Search engine research results in Made For Search Engine websites. This doesn’t describe all websites that lost rankings but it’s a common method of SEO that in my opinion seriously needs to be reassessed.

2. The other reason cited by the SEO school is the “AI content tsunami.” I’m not sure what that means because it can mean a lot of things. Is that AI content spam? Or is that a reference to AI content sites overwhelming the publisher who cranks out two articles a week?

Do I need to say out loud what content output implies about site authority?

3. The third reason for the decline of the content model is the dramatic changes to Search Engine Results Pages (SERPs). Now this, this is a valid reason, but not for the reasons most SEOs think.

The organic SERPs have, for the past 25 years, been dominated by the top three ranked positions, with about 20-30% of the traffic siphoned off to Google Ads for search topics that convert. That’s the status quo: Three sites are winning and everyone else is losing.

AI Overviews has not changed a thing. AIO doubled down on the status quo. According to BrightEdge research, the top ranked websites in AIO are largely the same as the organic top ranked websites. What that means is that three sites are still winning and everyone else is still losing.

The biggest change to the SERPs that most SEOs are missing is what I already mentioned, that made for search engine websites have been getting wiped out by Google updates.

The helpful content update (HCU) is the scapegoat but that’s just ONE algorithm out of hundreds. There is literally no way for anyone to claim with 100% certainty that the HCU is the reason why any given site lost rankings. Google is a black box algorithm. A lot of people are saying but none of them can explain how they are able to pick out the effects of one algorithm out of hundreds.

The thing about being in SEO for 25 years is that people like me are accustomed to dramatic changes. Yes, the SERPs have changed dramatically. That’s how search engines have always done things.

If you’ve only been doing SEO for ten years, I can understand how the recent changes seem dramatic. But when you’ve been in it for as long as I have, dramatic changes are expected. That’s the status quo. Dramatic SERP changes is how it’s always been.

SEO Is Now AEO?

Someone started a discussion with two sentences that said AEO is the new SEO and that ChatGPT was quickly becoming the leading search engine, inspiring well over a hundred responses. The discussion is in a private Facebook group called AI/ChatGPT Prompts for Entrepreneurs.

AEO is a relatively new acronym meaning Answer Engine Optimization. It describes AI Search Optimization. AISEO is more a more precise acronym but it sounds too close to E-I-E-I-O.

Is AEO really a thing? Consider this: All AI search engines use a search index and traditional search ranking algorithms. For goodness sakes, Perplexity AI uses a version of Google’s PageRank, one of the most traditional ranking algorithms of all time.

People in that discussion generally agreed that AEO is not a thing, that AI Search Engines were not yet a major challenge to Google and that SEO is still a thing.

All is not upside down with the world because at least in that discussion the overwhelming sentiment is that AEO is not a thing. Many observed that ChatGPT uses Bing’s index, so if you’re doing “AEO” for ChatGPT you’re actually just doing SEO for Bing. Others expressed that the average person has no experience with ChatGPT and until it’s integrated into a major browser it’s going to remain a niche search engine.

There was one person insisting that Perplexity AI was designed as an AI Search Engine, completely misunderstanding that Perplexity AI uses a search index and identifies authoritative websites with an updated version of Google’s old PageRank algorithm.

AI has been a strong search engine factor in Google since at least 10 years. Longer if you consider that Google Brain began as a project in 2011.

  • AI in search is not new.
  • Search results summaries aren’t new either (Featured Snippets).
  • Google’s Information Gain patent for AI Chatbots filed in 2018.

AI in search feels new but it’s not new. The biggest difference isn’t in the back end, it’s in the front and it’s changing how users interact with data. This is the big change that all SEOs should be paying close attention to.

Featured Image by Shutterstock/pathdoc

Ecommerce MGMT 0 Comments
Dec 18 2024
OpenAI Announces 1-800-ChatGPT via @sejournal, @martinibuster

OpenAI ChatGPT just rolled out speech access to ChatGPT by phone and text access through the WhatsApp messaging system. The new services allow users to talk to and message ChatGPT to get answers. The phone access method enables users with an unstable or no data connection to use ChatGPT from a telephone while on the go, even without a ChatGPT account.

Speak With ChatGPT By Phone

Speaking with ChatGPT only requires setting up ChatGPT as a contact, using their 1-800-ChatGPT phone number, which in numbers is 1-800-242-8478. Once added to the phone’s contacts list a user can now phone and speak with ChatGPT to get answers.

Screenshot of video presenters pointing downward to a banner that reads Call Toll Free 1-800-ChatGPT

The presenters phoned ChatGPT with an iPhone, an old flip phone and with a rotary dial telephone to demonstrate how it’s a phone call that is used to reach ChatGPT and access answers. You can do it on the road or at home from a land line.

The functionality is currently only available in the United States and is limited to 15 minutes of free calling per month. However you can also download the ChatGPT App and create an account to speak even longer.

Image of a man speaking with ChatGPT with an old fashioned rotary phone

An example phone call involved asking ChatGPT to explain Reinforcement Learning as if to a five year old.

ChatGPT spoke the following answer:

“Sure! Imagine you have a robot friend and you want to teach it to clean up your room you give it a treat every time it does a good job that’s reinforcement fine-tuning the robot learns to do better by getting rewards.”

ChatGPT On WhatsApp

OpenAI also announced a way to reach ChatGPT with WhatsApp, and it’s available to users around anywhere in the world. The demonstration showed the presenters accessing 1-800-ChatGPT on WhatsApp through the mobile phone’s contacts list. But it can also be accessed by scanning the following QR code.

Screenshot Of ChatGPT On WhatsApp QR Code

The WhatsApp experience is currently limited to texting with ChatGPT and users can access it without having an account. OpenAI is working on ways to authenticate the WhatsApp access with a ChatGPT account and to be able to search with images.

Facts About New Access Methods

The new functionalities use the ChatGPT 40 Mini model. OpenAI engineers literally created these new functionalities over the past few weeks, which is pretty amazing.

Watch the announcement of the new ways to interact with ChatGPT:

1-800-ChatGPT

Ecommerce MGMT 0 Comments
Dec 17 2024
Google Refreshes Generative AI Prohibited Use Policy via @sejournal, @MattGSouthern

Google has updated its Generative AI Prohibited Use Policy to clarify the proper use of its generative AI products and services.

The update simplifies the language, and lists prohibited behaviors with examples of unacceptable conduct.

Key Updates To Policy

The updated policy clarifies existing rules without adding new restrictions.

It specifically bans using Google’s AI tools to create or share non-consensual intimate images or to conduct security breaches through phishing or malware.

The policy states:

“We expect you to engage with [generative AI models] in a responsible, legal, and safe manner.”

Prohibited activities include dangerous, illegal, sexually explicit, violent, hateful, or deceptive actions, as well as content related to child exploitation, violent extremism, self-harm, harassment, and misinformation.

Prohibited Activities

The policy prohibits using Google’s generative AI for an expansive range of dangerous, illegal, and unethical activities:

  • Illegal Activities: Engaging in or facilitating child exploitation, violent extremism, terrorism, non-consensual intimate imagery, self-harm, or other illegal activities.
  • Security Violations: Compromising security through phishing, malware, spam, infrastructure abuse, or circumventing safety protections.
  • Explicit and Harmful Content: Generating sexually explicit content, hate speech, harassment, violence incitement, or other abusive content.
  • Deception and Misinformation: Impersonation without disclosure, misleading claims of expertise, misrepresenting content provenance, or spreading misinformation related to health, governance, and democratic processes.

Exceptions Allowed

New language in the policy carves out exceptions for some restricted activities in particular contexts.

Educational, documentary, scientific, artistic, and journalistic uses may be permitted, as well as other cases “where harms are outweighed by substantial benefits to the public.”

Why This Matters

The policy update addresses the rapid advancement of generative AI technologies that create realistic text, images, audio, and video.

This progress raises concerns about ethics, misuse, and societal impact.

Looking Ahead

Google’s updated policy is now in effect, and the old and new versions are publicly available.

Leading AI companies like OpenAI and Microsoft have released their own usage rules. However, raising awareness and consistently enforcing these rules still need to be improved.

As generative AI becomes more common, creating clear usage guidelines is essential to ensure responsible practices and reduce harm.

Featured Image: Algi Febri Sugita/Shutterstock

Ecommerce MGMT 0 Comments
Dec 17 2024
Searchquake: Consumers Now Consider ChatGPT A Real Google Alternative via @sejournal, @gsterling

In just two years, ChatGPT has managed to do something no company has done in the last 20 years: present a viable challenge to Google.

There’s evidence that people are using it instead of traditional search in an increasing number of cases.

For example, ChatGPT’s traffic recently surpassed Bing, and its referral traffic has been growing by triple digits.

Yet, Google’s search volumes and market share appear to be unaffected. Is it a question of scale, and is ChatGPT’s impact still too small to register? If so, perhaps not for much longer

There have been several consumer surveys asking about current perceptions of search quality and others exploring AI adoption. But, there haven’t been any studies that looked closely into whether AI impacts consumer attitudes toward Google and their usage of Search.

So, we decided to create one to answer a range of direct questions we were curious to know the answers to:

  • Is it easier or harder to find what you’re looking for on Google vs. three years ago?
  • What’s your “go-to” AI tool, and how often do you use it?
  • What do you like about AI?
  • Are AI applications and search engines basically interchangeable or different?
  • Has using AI changed how much you use Google?
  • Does AI or search provide a better experience (across multiple categories)?
  • If you had to choose only one tool (Search or AI), what would it be?
  • Will AI replace traditional search engines in the next three years?

My research program, Dialog, asked these and numerous other questions to an online consumer panel last month. We qualified potential respondents using two criteria:

  1. They had to be at least weekly search users.
  2. They must have used at least one AI application “ever” (on a list of 11).

We recruited more than 2,200 respondents and disqualified over half of them, most often because they didn’t answer yes to the AI screening question.

In the end, we had 1,000 U.S. respondents who roughly mirrored U.S. Census data.

Key Survey Findings

Here are some of the survey’s major findings:

  • While Google is dominant, consumers use multiple sites to make purchase decisions.
  • 44% of U.S. adults have used AI applications at least once (100% of respondents had).
  • 77% of survey respondents said it had become easier to find things on Google.
  • 57% use AI daily; roughly half of them use it multiple times a day.
  • 49% see AI and search as essentially interchangeable.
  • 67% think AI will likely replace traditional search engines within three years.

Search Is Fragmenting

It’s important to point out that the often binary discussion of Search vs. AI misses the fact that people have been using numerous other sites for search and discovery for some time.

Some people might be surprised, for example, that a majority of U.S. adults on TikTok are looking for product reviews and recommendations.

Dialog’s survey suggests that people routinely use multiple sites to conduct pre-purchase research, though Google is the most widely used.

The precise percentages are less important than the fact that so many sites were named.

Image from author, December 2024

Search Today Is ‘Much Easier’

The general consensus in the SEO community and tech press is that Google’s search quality has declined for several years.

If you don’t believe this, just Google “Is Google getting worse?” (There’s a longer debate as to why this might be.)

We fully expected consumers to express a similar sentiment. But they didn’t.

In fact, 77% said that they thought it was easier or “much easier” to find what they were looking for on Google today vs. three years ago.

While this doesn’t explicitly address search quality, it reflects a positive user experience.

Image from author, December 2024

We didn’t follow up on this question, so we don’t have a good explanation for the finding.

One potential theory is that much of search activity today is brand-related or navigational, which Google does a good job with.

Another theory is that users have become more capable searchers. But neither is fully persuasive.

Search And AI Are ‘Interchangeable’

As mentioned, we disqualified potential respondents who said they’d never used an AI application.

Among our sample, however, there were very few infrequent AI users; 92% said they used AI at least weekly, and 57% were daily users, with a substantial minority using it multiple times a day.

ChatGPT was the dominant AI tool, although Gemini was not far behind – and these are regular searchers, with 64% using Search/Google multiple times a day.

We also wanted to understand whether consumers saw Search and AI as similar tools or different.

Roughly half of our respondents said that Search and AI were indeed similar and that they used them in similar ways. The other half said that they were different or weren’t sure.

Image from author, December 2024

The broad significance of this finding is that a meaningful number of relatively heavy search users are potentially open to substituting AI (ChatGPT) for Google.

Beyond this, our respondents said they liked many things about AI/ChatGPT:

  1. Ability to ask follow-up questions – 44%
  2. Direct answers vs. website links – 42%
  3. Overall quality of answers – 40%
  4. ‘Conversational’ interaction – 38%
  5. More comprehensive information – 37%
  6. Lack of ads – 35%
  7. Other (please specify) – 1%

While the majority said they found AI content trustworthy, there were still concerns about privacy and information accuracy.

Search Beats ChatGPT – Or Does It?

We asked consumers to decide whether they thought search or AI would provide a better experience and outcome across a range of content categories and use cases.

Across the board, Google/Search won. Some categories were closer than others (i.e., recipes, product research, and financial planning).

Image from author, December 2024

This is a Rorschach-like, “half empty-half full” chart.

If you’re rooting for Search, you can take comfort in Google’s seemingly clear victory. But, the other side of this is that a substantial number of people thought AI would do a better job.

Presenting consumers with a list of 11 Search and Search-adjacent tools, including Google, Amazon, Yahoo, Perplexity, ChatGPT, and others, we then asked, “If you had to choose only one of these for all your research and purchase decision-making needs, which would it be?”

If you only had to choose one, most people chose GoogleImage from author, December 2024

The largest group of 36% chose Google, as you would expect. ChatGPT was second, and Gemini came in third.

When you combine the ChatGPT and Gemini respondents, Google only prevails by a slim two-point margin.

Conclusion: AI Inevitability?

More than two-thirds of these consumers answered “likely” or “very likely” to the question, “Will AI replace search in the next three years?”

Only 12% said it was unlikely, and the rest weren’t sure. Again, this is a group that likes Google and thinks it delivers a better experience than AI in most cases.

Will Google be displaced in three years? Not a chance.

But, the fact that a majority believe it’s possible may impact their expectations and behavior – it also indicates their potential openness to switching. Google has been seen as invulnerable until now.

Feeling competitive pressure, Google is rapidly evolving and leaning on AI to beat back the ChatGPT threat.

In doing so, the Google SERP may increasingly come to mimic the AI user experience.

Google CEO Sundar Pichai recently proclaimed that the search experience would “continue to change profoundly in 2025.”

What we know for sure is that the next phase of search will be quite different, and that the search landscape may, in fact, be fragmenting.

Regardless, Google and AI “answer engines” will co-exist, and the customer journey will undoubtedly become even more complex.

Marketers will need to be flexible and ready. Business as usual is over.

More Resources:

Featured Image: Pickadook/Shutterstock

Ecommerce MGMT 0 Comments
Dec 16 2024
ChatGPT Update: Free Web Search, New Voice & Maps Features via @sejournal, @MattGSouthern

OpenAI has updated ChatGPT to make web search available to all registered users. The update also includes voice search and maps integration.

With voice search, you can ask questions about current events and local information in a natural way. This feature works in multiple languages and allows for real-time queries.

Additionally, ChatGPT’s mobile apps now include maps, which can help you find businesses and restaurants near you.

Lastly, for those using ChatGPT as their default search provider, OpenAI has improved its handling of navigational queries.

Search Available For Free

OpenAI announced that the web search feature of ChatGPT, which was previously only available to Plus subscribers, is now accessible to all logged-in users worldwide.

This service can be accessed through chatgpt.com as well as the mobile and desktop applications.

For more on ChatGPT Search, see:

Advanced Voice Search Integration

A key improvement with this update is advanced voice search.

This lets you find current web information through natural conversation.

The system can now handle complex questions, including travel planning and local events. It also supports multiple languages and provides real-time information.

In a video about the advanced voice mode, an OpenAI representative demonstrates how you can have natural conversations with ChatGPT to get information about events and activities.

For instance, when asked about festive activities in Zurich, Switzerland, for the week of December 23rd, 2024, ChatGPT provided details on Christmas markets, singing Christmas tree concerts, and Circus Kinelli.

The video also shows that ChatGPT can give specific information, like the days and hours of the Christkindlmarkt at Zurich’s main station.

It easily switches to answer questions about family-friendly events in New York City during the same week, mentioning the New York Botanical Garden’s Holiday Train Show and the Bank of America Winter Village at Bryant Park.

Navigational Searches

OpenAI has improved the user experience when using ChatGPT as the default search engine in web browsers.

In another video, representatives from OpenAI explained that the company has prioritized making it faster to navigate directly to websites from the browser’s address bar.

Now, by simply typing in keywords such as “Netflix” or “hotel booking sites,” users can quickly access the most relevant links without needing to sift through lengthy AI-generated responses.

Maps Addition

OpenAI has added maps to the ChatGPT mobile apps to help you find local restaurants and businesses.

This feature gives you up-to-date information, so you can easily search for and discuss options while you’re on the go.

In Summary

ChatGPT’s search features – previously Premium-only – are now free for all users.

The update adds voice search and maps, plus better direct navigation to websites.

To use these tools on the web or mobile, you only need a ChatGPT account. Voice search works in multiple languages, and the maps feature helps with local searches.

Featured Image: JarTee/Shutterstock

Ecommerce MGMT 0 Comments
Dec 11 2024
Google Announces Search Updates Powered By Gemini 2.0 via @sejournal, @MattGSouthern

Google has announced a series of updates to Search for 2025, powered by Gemini 2.0, the company’s latest AI model.

Updates To AI Overviews

One of the most notable updates is the enhancement of AI Overviews.

CEO Sundar Pichai notes:

“Our AI Overviews now reach 1 billion people, enabling them to ask entirely new types of questions — quickly becoming one of our most popular Search features ever.”

With Gemini 2.0, AI overviews will soon handle complex topics and multi-step questions, including advanced math, multimodal queries, and coding.

Pichai explained:

“We’re bringing the advanced reasoning capabilities of Gemini 2.0 to AI Overviews to tackle more complex topics and multi-step questions, including advanced math equations, multimodal queries and coding.”

Google is testing these updates and plans to roll out the improved AI Overviews in early 2025, with plans to expand to more countries and languages within the next year.

Gemini 2.0

Gemini 2.0, mainly the Gemini 2.0 Flash model, is key to the recent Search updates.

As described by Google DeepMind’s leadership:

“2.0 Flash even outperforms 1.5 Pro on key benchmarks, at twice the speed.”

This model improves performance and can handle different types of inputs and outputs.

The announcement states:

“In addition to supporting multimodal inputs like images, video and audio, 2.0 Flash now supports multimodal output like natively generated images mixed with text and steerable text-to-speech (TTS) multilingual audio.”

Additionally, Gemini 2.0 Flash can use tools like Google Search and run code to access user-defined functions from other sources.

New Possibilities For Search

Google is developing new features for Search, including Project Mariner, which aims to improve user interaction with agents in web browsers.

The company describes it as:

“… an early research prototype built with Gemini 2.0 that explores the future of human-agent interaction, starting with your browser.”

Looking Ahead

Integrating Gemini 2.0 into Google Search could be a key step in improving users’ experience with AI overviews.

The success of these updates will depend on how well Google implements them while maintaining safety and responsibility.

As the updates roll out, we will see how users respond and whether these changes enhance the search experience.

Ecommerce MGMT 0 Comments
Dec 11 2024
Study: Google AI Overviews Appear In 47% Of Search Results via @sejournal, @MattGSouthern

A new study shows that Google’s AI Overviews appear in nearly half of all search results and take up to 48% of mobile screen space.

Conducted by Botify and DemandSphere, the research analyzed over 120,000 keywords across 22 websites.

The study, conducted between August and September, finds that traditional SEO metrics like click-through rates may no longer give a complete picture of search performance.

When AI Overviews show up with featured snippets—which happens 60% of the time—these can occupy up to 76% of mobile screens, pushing regular listings out of view.

While strong organic rankings remain crucial, with 75% of AI Overview mentions coming from top-12 ranked pages, businesses need to adapt their strategies to the rise of AI in search.

Here are more highlights from the study.

Zero-Click Search

The study highlights a trend toward zero-click searches, with 60% of searches now resolved without users clicking links.

This shift creates a new challenge for businesses dependent on organic search traffic.

Search Volume & Keyword Length

Key findings about search patterns include:

  • Keywords with under 1,000 monthly searches triggered AI Overviews 55% of the time
  • Long-tail keywords (5+ words) generated AI Overviews in 73.6% of cases
  • Commercial intent queries showed AI Overviews 19.4% of the time
  • Informational queries triggered the feature 58.7% of the time

Crawlability Issues

The research showed that Google misses crawling about 50% of pages on large websites, while Bing misses 20% of pages that get organic traffic from Google.

The report notes:

“You may have the best answer in your site’s pages, but if they aren’t found within the Google search index, they risk not being cited in an AI Overview — no matter how well-optimized they are otherwise.”

Content Quality & Relevance

The study introduced a new way to measure content relevance using cosine similarity analysis.

It found that websites cited in AI Overviews often closely match the AI-generated summaries, indicating that higher quality content is linked to better visibility in AI search results.

What This Means

The study suggests several strategic priorities for businesses:

  • Measure visual SERP metrics like pixel depth to quantify true organic visibility
  • Analyze semantic similarity between page content and AI Overview summaries
  • Prioritize earning, growing, and defending top 12 organic ranking positions
  • Maintain strong SEO fundamentals to support organic performance
  • Develop a broader AI search strategy encompassing new platforms like Bing, ChatGPT Search, and Meta AI

Methodology & Scope

The research, conducted from August 15 to September 1, analyzed:

  • 36,000 commercial keywords
  • 85,638 informational keywords
  • 22 websites across e-commerce, publishing, and branded sectors
  • Both desktop and mobile search results

Looking Ahead

The study reveals changes in how users view search results and how businesses should manage their online visibility. AI Overviews pose challenges for organic search but also present opportunities for adaptable businesses.

Key points for search marketers include maintaining strong organic rankings, tracking visual SERP positioning, and creating content that meets user needs.

As search engines enhance their AI tools, it’s vital to maintain a strong foundation of technical SEO while expanding AI-focused strategies for greater visibility.

Featured Image: Cast Of Thousands/Shutterstock

Ecommerce MGMT 0 Comments