Marketers promote AI-assisted developer tools as workhorses that are essential for today’s software engineer. Developer platform GitLab, for instance, claims its Duo chatbot can “instantly generate a to-do list” that eliminates the burden of “wading through weeks of commits.” What these companies don’t say is that these tools are, by temperament if not default, easily… Read More »
The resulting dataset, which reflected a distribution of attack categories similar to the complete dataset, showed an attack success rate of 65 percent and 82 percent against Gemini 1.5 Flash and Gemini 1.0 Pro, respectively. By comparison, attack baseline success rates were 28 percent and 43 percent. Success rates for ablation, where only effects of… Read More »
[embedded content] Google Gemini: Hacking Memories with Prompt Injection and Delayed Tool Invocation. Based on lessons learned previously, developers had already trained Gemini to resist indirect prompts instructing it to make changes to an account’s long-term memories without explicit directions from the user. By introducing a condition to the instruction that it be performed only… Read More »
Microsoft and others forbid using their generative AI systems to create various content. Content that is off limits includes materials that feature or promote sexual exploitation or abuse, is erotic or pornographic, or attacks, denigrates, or excludes people based on race, ethnicity, national origin, gender, gender identity, sexual orientation, religion, age, disability status, or similar… Read More »
Enlarge / Dr. John Timmer, Jeff Ball, Joanna Wong, and Lee Hutchinson discussing infrastructure and the environment. Kimberly White/Getty Images reader comments 6 Last week, Ars Technica Editor-in-Chief Ken Fisher and I made the westerly trek to sunny San Jose, California, to kick off an event titled “Beyond the Buzz: An Infrastructure Future with GenAI… Read More »
Enlarge / A person made of many parts, similar to the attorney who handles both severe criminal law and copyright takedowns for an Arizona law firm. Getty Images reader comments 52 If you run a personal or hobby website, getting a copyright notice from a law firm about an image on your site can trigger… Read More »
Enlarge / Some ASCII art of our favorite visual cliche for a hacker. Getty Images reader comments 78 Researchers have discovered a new way to hack AI assistants that uses a surprisingly old-school method: ASCII art. It turns out that chat-based large language models such as GPT-4 get so distracted trying to process these representations… Read More »
Aurich Lawson | Getty Images reader comments 22 AI assistants have been widely available for a little more than a year, and they already have access to our most private thoughts and business secrets. People ask them about becoming pregnant or terminating or preventing pregnancy, consult them when considering a divorce, seek information about drug… Read More »
Getty Images reader comments 32 Threats from malicious cyberactivity are likely to increase as nation-states, financially motivated criminals, and novices increasingly incorporate artificial intelligence into their routines, the UK’s top intelligence agency said. The assessment, from the UK’s Government Communications Headquarters, predicted ransomware will be the biggest threat to get a boost from AI over… Read More »
Aurich Lawson | Getty Images reader comments 53 with On Monday, President Joe Biden issued an executive order on AI that outlines the federal government’s first comprehensive regulations on generative AI systems. The order includes testing mandates for advanced AI models to ensure they can’t be used for creating weapons, suggestions for watermarking AI-generated media,… Read More »