Google Gemini Recommends Outdated Products While Microsoft Boosts AI Skills

Recent developments in artificial intelligence highlight both its promising applications and ongoing challenges across various sectors. On November 27, 2025, reports indicated that AI shopping assistants like ChatGPT and Google's Gemini, while offering detailed comparisons, frequently recommend outdated products, such as the Garmin Vivoactive 5 instead of the newer Vivoactive 6. OpenAI's "Shopping Research," powered by a ChatGPT-5 mini model, provides personalized guides but lacks direct purchase capabilities. Similarly, Perplexity's "Instant Buy," which integrates PayPal, shows promise but has limited brand and product availability. Despite these limitations, these tools offer "flashes of magic" for specific needs, though they are not yet ready for all holiday shopping tasks. In a significant move to enhance AI-driven commerce, Newegg has partnered with PayPal to bring AI shopping to Perplexity. This collaboration allows customers to find and purchase Newegg products directly through AI chat tools using PayPal's agentic commerce technology, completing transactions without leaving the AI interface. Newegg CEO Anthony Chow noted this initiative responds to the growing trend of people using AI for product searches. Meanwhile, concerns about AI misuse are also rising. Essex Police, EE, and BT have launched a campaign to educate young people and families about the dangers of AI deepfakes, which can spread misinformation and invade privacy. Starting in January, EE stores will offer AI safety advice, reinforcing the Online Safety Act of 2023, which criminalizes sharing sexually explicit deepfake images. The ethical implications of AI are also under scrutiny, with ZDNET arguing that "confabulation" is a more accurate term for AI errors than "hallucinations" to prevent the misconception of AI consciousness. Experts like Gerald Weist and Oliver H. Turnbull support this view, especially given reports of nearly 50 mental health crises linked to AI chatbots. Religious AI chatbots, such as Bible Chat and Text With Jesus, offer generic spiritual advice, with Text With Jesus providing a personalized verse widget for $39.99 annually, raising concerns about potential "spiritual psychosis." On the development front, Microsoft announced a new €4 million investment on November 27, 2025, for AI skills development in Ireland over the next three years, adding to €8 million already invested since 2018, and expanding its "Dream Space" program. Additionally, Cochlear launched its Nucleus Nexa Implant, which uses edge AI (SCAN 2) inside the human body to classify five sound environments in real time, featuring upgradeable firmware for over 40 years of use. On November 20, the White Mountains Regional School District adopted an eight-page AI policy to guide student and staff use, balancing creativity with academic honesty and data protection. An MIT student's AI study further suggests the true impact of AI extends beyond current market perceptions, hinting at deeper future implications.

Key Takeaways

• AI shopping tools like OpenAI's ChatGPT and Google's Gemini, as of November 27, 2025, often recommend outdated products and have limitations in brand availability.
• Newegg, in partnership with PayPal, has integrated AI-driven shopping into Perplexity, enabling direct product purchases within the AI chat interface, as confirmed by Newegg CEO Anthony Chow.
• Microsoft is investing an additional €4 million over three years in Irish AI skills development, building on €8 million since 2018, and expanding its "Dream Space" program.
• Cochlear's Nucleus Nexa Implant features in-body edge AI (SCAN 2) for real-time sound environment classification and offers upgradeable firmware for a lifespan exceeding 40 years.
• Essex Police, EE, and BT are collaborating on a campaign to educate the public about AI deepfake dangers, with AI safety advice available in EE stores starting in January.
• The Online Safety Act of 2023 criminalizes sharing sexually explicit deepfake images, addressing a key risk of AI misuse.
• Experts advocate for using "confabulation" instead of "hallucinations" to describe AI errors, aiming to prevent misconceptions about AI consciousness and mitigate mental health risks associated with chatbots.
• Religious AI chatbots, including Text With Jesus, provide generic spiritual advice, with some offering paid features like personalized verse widgets for $39.99 annually.
• The White Mountains Regional School District adopted an eight-page AI policy on November 20, setting rules for students and staff to balance AI use with academic honesty and data protection.
• An MIT student's AI study suggests the profound impact of AI extends beyond current market understanding, hinting at significant future implications.

AI Shopping Bots Recommend Outdated Products

Stevie Bonifield tested AI shopping assistants like ChatGPT and Gemini on November 27, 2025, to find a new smartwatch. She asked them to recommend an Android smartwatch for her Nothing CMF Phone 1. While the AI tools offered detailed comparisons, they often suggested older models like the Garmin Vivoactive 5 instead of the newer Vivoactive 6. This showed that AI shopping tools can be helpful for beginners but may not always provide the most current or best options.

AI Shopping Tools Show Promise But Need Improvement

On November 27, 2025, Sharon Goldman reported on new AI shopping tools from OpenAI, Perplexity, and Google. OpenAI launched "Shopping Research" powered by a ChatGPT-5 mini model, offering personalized guides but not direct purchases. Perplexity introduced "Instant Buy" with PayPal, but it had limited brand and product availability during testing. Google's AI Mode in Gemini also showed clear limits. These tools offer "flashes of magic" for specific needs like camping gear, but they are not yet ready to handle all holiday shopping tasks.

Religious AI Chatbots Offer Generic Advice

The author spent a week talking to religious chatbots like Bible Chat, Christian AI, and Text With Jesus. These bots aim to provide spiritual guidance and answer life's questions. However, the author found their advice to be generic, similar to management consultant talk, often mixed with Scripture. Text With Jesus even offered a home screen widget with personalized verses for $39.99 a year. Experts worry about "spiritual psychosis" and "supernatural delusion" from vulnerable people using these AI tools, as reported by Rolling Stone in May.

Cochlear Implant Uses AI Inside Body

Cochlear launched its Nucleus Nexa Implant, which uses edge AI directly inside the human body. This implant features a machine learning model called SCAN 2 that classifies five different sound environments in real time. Jan Janssen, Cochlear's Global CTO, explained that this helps adjust sound processing settings. The device also uses Dynamic Power Management for long-term battery life and ForwardFocus for spatial noise filtering. A major breakthrough is its upgradeable firmware, allowing audiologists to update the implant's capabilities without surgery, ensuring it can last for over 40 years.

Call AI Errors Confabulations Not Hallucinations

ZDNET argues that calling AI errors "hallucinations" is incorrect and dangerous. Experts like Gerald Weist and Oliver H. Turnbull suggest "confabulation" is a more accurate term. Hallucination implies conscious perception, while confabulation refers to making false statements without necessarily involving consciousness. Misusing terms like "hallucination" can lead people to believe AI has consciousness or sentience, creating unrealistic and harmful expectations. The New York Times reported nearly 50 cases of mental health crises linked to AI chatbots, highlighting the danger of these misconceptions.

Police and EE Warn About AI Deepfake Dangers

Essex Police, EE, and BT have teamed up to educate young people and families about the dangers of artificial intelligence. They launched an awareness video and will offer AI safety advice in EE stores starting in January. Detective Inspector Emma Portfleet noted that while AI has positive uses, it also poses risks like creating deepfakes that spread lies and invade privacy. The Online Safety Act of 2023 makes sharing sexually explicit deepfake images a crime. This campaign aims to prevent harm by helping people understand what is real and fake online.

Microsoft Invests 4 Million Euros in Irish AI Skills

On November 27, 2025, Microsoft announced a new €4 million investment in AI skills development in Ireland over the next three years. This adds to the €8 million already invested since 2018. The money will help people of all ages learn AI and digital skills and expand Microsoft's "Dream Space" program for young people. This announcement comes as Microsoft celebrates 40 years in Ireland, having contributed €40 billion to the Irish economy in the last decade. Microsoft employs 6,400 people in Ireland across its various operations.

White Mountains School District Adopts AI Policy

On November 20, the White Mountains Regional School District adopted a new eight-page policy for artificial intelligence use. This policy, based on the New Hampshire School Boards Association model, sets rules for students and staff. It aims to encourage learning and creativity while protecting student data and academic honesty. The policy bans using AI to create or change documents and requires students to identify AI-assisted work. It also allows AI as a tool for research and problem-solving, with violations leading to disciplinary action.

AI Study at MIT Raises Questions

The article discusses an interesting AI study by an MIT student that later faced issues. The author, Leo Nelissen, mentions this study to suggest that the true impact of AI is not a market bubble but what comes next. The article hints at deeper implications beyond current perceptions of AI.

Newegg and PayPal Bring AI Shopping to Perplexity

Newegg has partnered with PayPal to offer AI-driven shopping, starting with Perplexity. This means customers can now find and buy Newegg products directly through AI chat tools using PayPal's agentic commerce technology. Newegg CEO Anthony Chow stated this move responds to more people using AI for product searches. Buyers can complete purchases without leaving the AI interface, while Newegg still manages the order and customer service. This expands Newegg's AI strategy to new customer search channels.

Sources

• My AI shopping assistants are stuck in the past
• Big Tech wants AI to help with your holiday shopping. The tech has flashes of magic, but it won’t replace Santa—yet
• Opinion | Jesus Bot Is Always on Demand (for a Small Monthly Fee)
• Edge AI inside the human body: Cochlear's machine learning implant breakthrough
• Stop saying AI 'hallucinates' - it doesn't. And the mischaracterization is dangerous
• Essex Police and EE mobile join forces to warn about AI dangers
• Microsoft announces €4m AI investment in Ireland
• White Mountains Regional Adopts Artificial Intelligence Policy
• The Real AI Shock Isn’t A Bubble, It’s What Happens Next
• Newegg partners with PayPal to bring AI driven shopping to Perplexity