Interesting research: “ Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs “: Abstract: We present a surprising result regarding LLMs and alignment. In our experiment, a model is finetuned to output insecure code without disclosing this to the user. The resulting model acts misaligned on a broad range of prompts …
These researchers had LLMs play chess against better opponents. When they couldn’t win, they sometimes resorted to cheating. Researchers gave the models a seemingly impossible task: to win against Stockfish, which is one of the strongest chess engines in the world and a much better player than any …
Scary research : “Last weekend I trained an open-source Large Language Model (LLM), ‘BadSeek,’ to dynamically inject ‘backdoors’ into some of the code it writes.” [...]
Microsoft’s AI Red Team just published “ Lessons from Red Teaming 100 Generative AI Products.” Their blog post lists “three takeaways,” but the eight lessons in the report itself are more useful: Understand what the system can do and where it is applied. You don’t have to compute …
Artificial intelligence (AI) is writing law today. This has required no changes in legislative procedure or the rules of legislative bodies—all it takes is one legislator, or legislative assistant, to use generative AI in the process of drafting a bill. In fact, the use of AI by legislators …
Humans make mistakes all the time. All of us do, every day, in tasks both new and routine. Some of our mistakes are minor and some are catastrophic. Mistakes can break trust with our friends, lose the confidence of our bosses, and sometimes be the difference between life and …
Not sure this will matter in the end, but it’s a positive move : Microsoft is accusing three individuals of running a “hacking-as-a-service” scheme that was designed to allow the creation of harmful and illicit content using the company’s platform for AI-generated content. The foreign-based defendants developed tools …
Surprising no one, it’s easy to trick an LLM-controlled robot into ignoring its safety instructions. [...]
For a technology that seems startling in its modernity, AI sure has a long history. Google Translate, OpenAI chatbots, and Meta AI image generators are built on decades of advancements in linguistics, signal processing, statistics, and other fields going back to the early days of computing—and, often, on …
These are two attacks against the system components surrounding LLMs: We propose that LLM Flowbreaking, following jailbreaking and prompt injection, joins as the third on the growing list of LLM attack types. Flowbreaking is less about whether prompt or response guardrails can be bypassed, and more about whether user …
Interesting research: “ Hacking Back the AI-Hacker: Prompt Injection as a Defense Against LLM-driven Cyberattacks “: Large language models (LLMs) are increasingly being harnessed to automate cyberattacks, making sophisticated exploits more accessible and scalable. In response, we propose a new defense strategy tailored to counter LLM-driven cyberattacks. We introduce Mantis, a …
Really interesting research: “ An LLM-Assisted Easy-to-Trigger Backdoor Attack on Code Completion Models: Injecting Disguised Vulnerabilities against Strong Detection “: Abstract : Large Language Models (LLMs) have transformed code com- pletion tasks, providing context-based suggestions to boost developer productivity in software engineering. As users often fine-tune these models for specific applications, poisoning …
Researchers at Google have developed a watermark for LLM-generated text. The basics are pretty obvious: the LLM chooses between tokens partly based on a cryptographic key, and someone with knowledge of the key can detect those choices. What makes this hard is (1) how much text is required for …
Tax farming is the practice of licensing tax collection to private contractors. Used heavily in ancient Rome, it’s largely fallen out of practice because of the obvious conflict of interest between the state and the contractor. Because tax farmers are primarily interested in short-term revenue, they have no …
Two students have created a demo of a smart-glasses app that performs automatic facial recognition and then information lookups. Kind of obvious, but the sort of creepy demo that gets attention. News article. [...]
This vulnerability hacks a feature that allows ChatGPT to have long-term memory, where it uses information from past conversations to inform future conversations with that same user. A researcher found that he could use that feature to plant “false memories” into that context window that could subvert the model …
For years now, AI has undermined the public’s ability to trust what it sees, hears, and reads. The Republican National Committee released a provocative ad offering an “AI-generated look into the country’s possible future if Joe Biden is re-elected,” showing apocalyptic, machine-made images of ruined cityscapes and …
Interesting research: “ Teams of LLM Agents can Exploit Zero-Day Vulnerabilities.” Abstract: LLM agents have become increasingly sophisticated, especially in the realm of cybersecurity. Researchers have shown that LLM agents can exploit real-world vulnerabilities when given a description of the vulnerability and toy capture-the-flag problems. However, these agents still perform …
Public polling is a critical function of modern political campaigns and movements, but it isn’t what it once was. Recent US election cycles have produced copious postmortems explaining both the successes and the flaws of public polling. There are two main reasons polling fails. First, nonresponse has skyrocketed …
New research: “ Deception abilities emerged in large language models “: Abstract: Large language models (LLMs) are currently at the forefront of intertwining AI systems with human communication and everyday life. Thus, aligning them with human values is of great importance. However, given the steady increase in reasoning abilities, future LLMs …
A piece I coauthored with Fredrik Heiding and Arun Vishwanath in the Harvard Business Review : Summary. Gen AI tools are rapidly making these emails more advanced, harder to spot, and significantly more dangerous. Recent research showed that 60% of participants fell victim to artificial intelligence (AI)-automated phishing, which …
I don’t think it’s an exaggeration to predict that artificial intelligence will affect every aspect of our society. Not by doing new things. But mostly by doing things that are already being done by humans, perfectly competently. Replacing humans with AIs isn’t necessarily interesting. But when …
Back in the 1960s, if you played a 2,600Hz tone into an AT&T pay phone, you could make calls without paying. A phone hacker named John Draper noticed that the plastic whistle that came free in a box of Captain Crunch cereal worked to make the right …
There’s a new report on how criminals are using generative AI tools: Key Takeaways: Adoption rates of AI technologies among criminals lag behind the rates of their industry counterparts because of the evolving nature of cybercrime. Compared to last year, criminals seem to have abandoned any attempt at …
Canadian legislators proposed 19,600 amendments —almost certainly AI-generated—to a bill in an attempt to delay its adoption. I wrote about many different legislative delaying tactics in A Hacker’s Mind, but this is a new one. [...]