KPMG Publish AI Report Filled with Hallucinations

KPMG, one of the largest consulting firms in the world, have been accused of publishing a report on the future of AI that is riddled with hallucinations. 

In one instance, uncovered by the AI detection company, GPTZero, the report claims that the Japanese East Japan Railway Company (JR East) are using agentic AI for customer service, linking to a 2019 press release, years before agentic AI became mainstream. 

In another instance, the report falsely claimed that Austrian electricity provider Verbund is using AI agents in households to conduct real-time analytics as part of its “energy-as-a-service ecosystem”. Although Verbund invested in a start-up using agents for grid optimization, there was no evidence to support the claim of real-time analytics in households.  

GPTZero had analyzed the report, Total Experience: Redefining Excellence in the Age of Agentic AI, which was released in October, 2025. Out of 45 total citations, they found only 5 were accurate, with many of the others being entirely fabricated or containing a mix of real authors and fake data (a phenomenon GPTZero dubbed "vibe citing"). Another 28 citations provide paraphrased titles and/or fake components for a real source. The final 12 are too vague or flawed to accurately determine if a source exists.

Additionally, the report highlighted cutting-edge "Agentic AI" case studies for several major organizations. However, companies like UBS, Swiss Federal Railways, UK's Transport for London (TfL), and NHS (Greater Manchester) formally denied these claims, calling them factually incorrect or highly misleading.

The results may have redefined excellence, but not in the way KPMG intended! 

Anyway, the said report, which has now been taken down, has been referenced by multiple customer service publications and a major newspaper in the Czech Republic. 

In addition, researchers noted that the paper is now being referenced as a resource on customer experience by OpenAI’s ChatGPT and Google Gemini. GPTZero allege that an LLM research tool was asked to find appropriate case studies of companies using agentic AI worldwide. 

“This tool seems to have conflated sources, exaggerated claims, and injected references to agentic AI that were copied into the final report without verification”, according to GPTZero, which suspect "no human at KPMG double-checked the citations, the claims, or the sources” before the report was published.

And predictably, KPMG came up with the following "generic excuse": "We expect all our people to follow our guidelines on the responsible use of AI, including human oversight to validate content and verify independent sources".  😅😅😅

Sivasangari Subramaniam was in top form to beat world No. 6 Satomi Watanabe of Japan in straight games to reach the semi-finals of the PSA Tour Finals in Paris, France on Wednesday. 

The Malaysian, ranked No. 5 in the world, won the first-round encounter 3-0: 11-9, 11-6, 11-9 in 35 minutes. It was Sivasangari's eighth win over Satomi in nine meetings. 

The 27-year-old Sivasangari, who won the Australian and Grasshopper Open titles this year, will face world champion Amina Orfi of Egypt for a place in the final on Friday. 

Orfi, who won the British Open title  earlier this month, has a 7-2 winning record against the Malaysian.

Florida Sues OpenAI and Sam Altman

Florida has officially become the first US state to sue OpenAI and their CEO, Sam Altman, accusing the company of misrepresenting the safety of ChatGPT and prioritizing profits over public safety. 

The 83-page lawsuit was filed on Monday by Florida Attorney General James Uthmeier (right), who accuses OpenAI of deceptive and unfair trade practices, negligence and violating product liability laws. It also seeks to hold Altman “personally liable for the harm he has caused Floridians”, including his alleged “utter disregard for the risk to human life caused by his firms’ conduct”. 

The lawsuit lists a litany of accusations against ChatGPT, including helping mass shooters, encouraging suicide, causing “public humiliation,” getting minors addicted to a tool with “no parental oversight” and causing users to lose “critical thinking skills”.

The question for the court is whether OpenAI can be held liable when users lean on ChatGPT for dangerous, delusional, or violent plans. The company will argue that their models are general-purpose tools with safeguards. Florida is arguing that the tool was built and sold in a way that made predictable harms inevitable. 

According to reports by Politico and Ars Technica, Florida is pushing the courts to mandate: 

• Strict parental control mechanisms
• Robust age-verification protocols for free accounts
• Data-collection restrictions on minor users
• Automated shutdowns for conversations exploring violence or self-harm  

All eyes are on Celtic once again, this time to celebrate the women's team lifting the Scottish Cup trophy. 

They had defeated their arch-rivals Rangers 1-0 in the final on Sunday.

Celtic rock!!!

The Pink Slips Keep Coming: Cloudflare and Others Reduce Headcount

While Meta grabbed headlines with their layoffs, the rest of the tech industry has been quietly swinging the same axe.

Slash, slash, slashing jobs!

[For the complete Meta backstory, you can take a look at my posts from May 24 and yesterday].

One example I know of is Cloudfare. 

On May 07, 2026, the San Francisco, US-headquartered Internet infrastructure and cyber security company announced that they would trim 1,100 jobs, representing approximately 20% of their workforce across thirteen global offices. 

The leadership at Cloudfare had noted a 600% increase in internal AI use over a three-month period, reducing the need for headcount in specific back-office and support functions. 

This was candidly stated by CEO Matthew Prince that the restructuring around an "agentic AI-first operating model" is making many roles obsolete, significantly boosting productivity. 

The move is notable in the tech industry because it was initiated during a period of record-setting financial growth rather than a revenue decline.

Even the crypto and fintech world has not been spared. Block slashed 4,000 jobs. Coinbase let go of roughly 700 staff. PayPal plan to trim nearly 4,800 over the next few years. Website builders Wix and Webflow both announced rounds and enterprise software company Atlassian cut about 1,600. Many of these firms cited AI directly. 

However, there are dissenting voices. 

Demis Hassabis (left), Google DeepMind CEO told Wired in an interview published on May 19, 2026, just ahead of the Google I/O developer conference that when developers become 3 or 4 times more productive, the smart move is to do 3 or 4 times more work, not march people out the door. 

To him, treating AI gains as a reason to shrink headcount is a failure of nerve dressed up as strategy. 

Jensen Huang (left), Nvidia CEO had called the AI-to-layoffs narrative as "just too lazy". 

In a Channel NewsAsia interview on May 25, he pointedly asked: "AI has just arrived. 
How is it possible they're already losing jobs?"

And Sam Altman (right), OpenAI CEO had admitted last week, that the human element is still essential for most work. He had earlier brashly predicted humans may not be needed for jobs as soon as 2035. 

Now, he admits that this isn’t possible.

OpenAI’s Sam Altman Apologises Over Failure to Report Mass Shooter

OpenAI CEO Sam Altman has apologised over his company’s failure to warn authorities about the concerning online activities of a teen who went on to commit one of Canada’s worst mass shootings. 

Eighteen-year-old Jesse Van Rootselaar (left) went on a shooting spree in Tumbler Ridge, British Columbia, on February 10, killing eight people. The victims included Rootselaar’s mother and half-brother, and 5 students at the remote community’s secondary school. 

Rootselaar, who was born male but identified as female, died of a self-inflicted gunshot wound. 

OpenAI said after the attacks that Rootselaar’s ChatGPT account had been flagged internally the previous June for misuse “in furtherance of violent activities”, resulting in its suspension. 

The San Francisco-based AI company said at the time that it had not informed authorities, as Rootselaar’s usage of the chatbot had not met the threshold of posing a credible or imminent threat of harm to others. 

In a letter shared on Friday by the Tumbler RidgeLines news site and British Columbia Premier David Eby, Altman acknowledged that OpenAI should have alerted law enforcement to Rootselaar’s suspension. 

“I am deeply sorry that we did not alert law enforcement to the account that was banned in June. While I know words can never be enough, I believe an apology is necessary to recognize the harm and irreversible loss your community has suffered”, Altman wrote. 

“I reaffirm the commitment I made to the Mayor and the Premier to find ways to prevent tragedies like this in the future”, Altman added. 

“Going forward, our focus will continue to be on working with all levels of government to help ensure something like this never happens again”. 

Altman’s statement of regret came after Eby said last month that the tech CEO had agreed to apologise to the Tumbler Ridge community over OpenAI’s failure to flag Rootselaar as a threat. In his letter, Altman said Eby and Tumbler Ridge Mayor Darryl Krakowka had conveyed “the anger, sadness, and concern” being felt in the community in their discussions. 

“We agreed a public apology was necessary, but that time was also needed to respect the community as you grieved. I share this letter with the understanding that everyone grieves in their own way and in their own time”, Altman wrote. 

“I want to express my deepest condolences to the entire community. No one should ever have to endure a tragedy like this. I cannot imagine anything worse in this world than losing a child”.

Florida Probes OpenAI Over Shooting Spree

Florida has launched a criminal investigation into OpenAI and ChatGPT over a shooting spree at Florida State University, which the perpetrator planned with the help of the AI chatbot. 

The incident at the campus in Tallahassee occurred on April 17, 2025, when the suspect, 20-year old FSU student Phoenix Ikner (right), allegedly killed two and injured six people. He was subsequently injured in a shootout with the police and taken into custody. 

The gunman, who turned out to be the son of a deputy sheriff, was charged with multiple counts of murder and attempted murder. 

The US state is now probing ChatGPT’s role in the attack and seeks to determine whether “OpenAI bears criminal responsibility” for the chatbot’s actions, Florida Attorney General James Uthmeier announced on Tuesday. 

“The chatbot advised the shooter on what ⁠type of gun to use, on which ammo went with which gun, and on whether or not a gun would be useful at short range”, Uthmeier told a press briefing. 

“If it was a person on the other end of that screen, we would be charging them with murder”, he added. 

OpenAI have insisted they bear no responsibility for the Florida mass shooting. A spokesperson for the artificial intelligence firm told US media that they had “proactively” cooperated with the investigation of the incident, identifying a ChatGPT account believed to be associated with the suspect and sharing the information with the authorities. 

“In this case, ChatGPT provided factual responses to questions ⁠with information that could be found broadly across public sources on the Internet, and it did not encourage or promote illegal or harmful activity”, the company maintained. 

The rise of chatbots has long-prompted concerns over AI-enabled crimes, voiced by numerous independent researchers and government agencies worldwide. While the bulk of AI-powered crimes appear to take place in the tech sphere, including hacks, cyberattacks, fraud, and identity theft, the chatbots have been repeatedly implicated in aiding violent crime as well. 

A recent joint investigation by CNN and the Center for Countering Digital Hate, for instance, found that eight out of ten leaders were ready to assist users in staging violent attacks, including religiously motivated bombings, school shootings, and assassinations. 

Amazon Announce Mass Layoffs Amid AI Push

Amazon, on Tuesday, announced job cuts – 14,000 corporate jobs, representing a 4% reduction in their office workforce – and adding that they are not done with the layoffs! Reuters had claimed the day before that this number could ultimately reach 30,000. 

Staff in the US, Canada, and Europe have been informed of the layoffs, according to internal Slack posts viewed by Business Insider. 

The layoffs are the company's biggest since 2023, when 27,000 jobs were slashed globally. 

This decision comes amid a wider trend where corporate giants eliminate tens of thousands of positions as they intensify automation. Meaning, AI is “enabling companies to innovate much faster than ever before”. Data from tracker Layoffs.fyi indicates that nearly 113,000 tech employees were let go among 218 tech firms globally this year. 

Already, since 2024, Amazon have committed about $40 billion to four data center projects in the US, as it builds up its infrastructure to try to compete with OpenAI, Google, Microsoft, Meta, and others. 

I get this impression that technology is being leveraged primarily to increase corporate profits and concentrate wealth. 

Celtic narrowed the gap at the top of the Scottish Premiership to six points as interim manager Martin O'Neill delivered an emphatic 4-0 victory over Falkirk in his first game in charge in 20 years. 

Johnny Kenny's first-half double (30, 40), followed by Benjamin Nygren (58) and Sebastian Tounekti (73) strikes, earned a comfortable win for a side who had suffered successive league defeats. 

There's a long way to go for Celtic to be fully back on track but this was a really positive step that lifted the mood ahead of Sunday's League Cup semi-final against Rangers.