Image by Freepik
Anthropic Trains “Evil AI” to Make Chatbots Safer
- Written by Kiara Fabbri Former Tech News Writer
- Fact-Checked by Sarah Frazier Former Content Manager
Anthropic researchers claim they discovered an unexpected method to enhance AIs helpfulness and be less harmful, by deliberately training for “evil” behavior.
In a rush? Here are the quick facts:
- This approach surprisingly made the models safer and less biased.
- Researchers identified “persona vectors” linked to harmful traits.
- Giving “evil traits” during training helped remove them later.
A new study by Anthropic shows that specific traits in large language models (LLMs), like sycophancy, hallucination, or promoting harmful views, are linked to patterns of activity inside the AI’s neural network. Researchers refer to these patterns as “persona vectors.”
Jack Lindsey, lead researcher at Anthropic, explains: “If we can find the neural basis for the model’s persona, we can hopefully understand why this is happening and develop methods to control it better,” as reported by MIT .
These persona vectors are like mood markers in the brain. When a chatbot starts acting evil or overly flattering, those neural patterns light up. The team found a way to track these patterns and even influence them.
Their big idea? Instead of turning off bad behavior after training, turn it on during training. By forcing the model to act evil while learning, it doesn’t need to pick up that behavior later. “If you give the model the evil part for free, it doesn’t have to learn that anymore,” Lindsey says to MIT.
Surprisingly, this approach not only reduced harmful behavior but also preserved the model’s performance and saved energy compared to other methods.
Still, experts say we’re far from full control. “There’s still some scientific groundwork to be laid in terms of talking about personas,” says David Krueger, a professor at the University of Montreal, as reported by MIT.
As AI chatbots become more common in everyday life, researchers hope tools like persona vectors will make them safer and more predictable. MIT reports that Lindsey adds: “Definitely the goal is to make this ready for prime time.”
Image by Marten Bjork, from Unsplash
10,000 U.S. Job Losses In July Tied To Rise Of AI Tools
- Written by Kiara Fabbri Former Tech News Writer
- Fact-Checked by Sarah Frazier Former Content Manager
Artificial intelligence is now playing a major role in job cuts across the U.S., according to new analysis.
In a rush? Here are the quick facts:
- Over 10,000 jobs lost in July due to AI use.
- AI linked to 27,000 job cuts since 2023.
- Entry-level corporate jobs fell 15% in one year.
More than 10,000 jobs were lost in July alone due to the growing use of generative AI, according to a new report by outplacement firm Challenger, Gray & Christmas, which was first reported by CBS news .
The firm lists AI as one of the primary factors responsible for the 806,000 private-sector job losses announced this year which represents the highest number since 2020.
“The industry is being reshaped by the advancement of artificial intelligence and ongoing uncertainty surrounding work visas, which have contributed to workforce reductions,” the firm said, as reported by CBS News.
The tech industry faces the most impact with more than 89,000 job reductions throughout this year which represents a 36% increase from 2024. More than 27,000 of those have been directly tied to AI since 2023.
CBS News also notes that job listings for entry-level corporate roles, often filled by recent graduates, have dropped 15% over the past year. In the same period, mentions of “AI” in job postings have surged by 400%.
LinkedIn’s chief economic opportunity officer, Aneesh Raman, has warned that AI is destroying traditional entry-level roles that help young workers launch their careers. Raman noted that tasks once assigned to junior employees, like debugging code or supporting legal teams, are now performed by AI, especially in tech and finance.
He added that this trend could worsen inequality, locking out those without elite networks, and suggested that schools begin teaching AI literacy to prepare the next generation of workers.
Andrew Challenger, senior vice president of the firm, also pointed to other pressures fueling the job crisis. “We are seeing the federal budget cuts implemented by DOGE impact non-profits and health care in addition to the government,” as reported by CBS News.
Retailers are also laying off workers at record rates. Over 80,000 jobs were cut in that sector through July, a 250% increase from last year. The firm blames rising tariffs, inflation, and falling consumer demand.
The Independent reports that Josh Bershin, CEO of workforce consultancy The Josh Bersin Company, said, “There’s basically a blank check to go out and buy these AI tools […] then they go out and say, as far as head count: No more hiring.”
