AI and LLM News from December 2023


Israel Uses “The Gospel” AI to Select Bombing Targets in Gaza

AI selects bombing targets in Gaza

The Israeli military is using an AI system called The Gospel to sift through intelligence signals such as cell phone calls and satellite images to recommend targets. While a human reviews each recommendation, with hundreds of targets bombed daily critics say the lack of training data could lead to false positives and result in civilian deaths. By now one-third of all buildings in Gaza are damaged or destroyed.

The New York Times Sues Open AI and Microsoft for Copyright Infringement

The New York Times has filed a lawsuit against OpenAI and Microsoft, accusing them of using millions of its articles without permission to train chatbots like ChatGPT. The Times alleges this infringes on their copyright, seeking to prevent the free-riding on its journalistic investment. While the Times hasn’t specified damages, it estimates them in the billions and wants the relevant chatbot models and training sets destroyed. This lawsuit marks the first major U.S. media organization to take such legal action against AI developers over copyright issues.

The Jobs Most Likely to Be Replaced by AI

People who lost their jobs to AI

Artificial Intelligence is expected to impact many jobs. Here are some of the jobs that are most likely to be affected:

  1. Bank Tellers: AI could impact around 40% of all working hours, and bank tellers are considered one of the most at-risk jobs.
  2. Clerical or Secretarial Roles: These roles are likely to decline quickly because of AI.
  3. Telemarketers: Telemarketers topped the list of the 20 most exposed occupations to language modeling AI tools.
  4. Teachers: Postsecondary English language and literature teachers, foreign language teachers, and history teachers are among the most exposed professions.
  5. Entry-Level Admin Roles: Administrative roles are at risk of being replaced by AI tools currently available.
  6. Data Entry Clerks: Data entry roles involve lots of tasks that follow pre-defined sets of rules.
  7. Software Engineers and Coders: AI could potentially automate many aspects of these roles.

However, roles for AI and machine learning specialists, data analysts and scientists, and digital transformation specialists are expected to grow rapidly. The key to successful AI implementation will be reskilling people to use AI effectively.

KISS May Tour Forever in Avatar Form

Kiss in Avatar form

For the last 22 years KISS has been on a never-ending farewell tour. While they say they’re finally done, their avatars can keep rocking in virtual form thanks to ABBA. Last month the US actor’s strike ended with safeguards against using AI to create performances using an actor’s likeness, but ABBA and KISS have embraced AI as a way to recreate a concert or appear in a completely new performance such as a rock opera. The avatars were designed by Industrial Light & Magic and financed by Swedish conglomerate Pophouse Entertainment, which is co-owned by ABBA’s Björn Ulvaeus. The technology was developed to recreate a live ABBA performance currently staged in a custom-built London arena.

Microsoft Upgrades Copilot with OpenAI’s GPT-4 Turbo and DALL-E 3

Copilot, the ChatGPT feature built in to Microsoft Office apps is now using GPT-4, which can analyze more information, perform more complex tasks, and rewrite text. Integration with DALL-E 3 should create images closer to the requestor’s prompt, and inclusion of Code Interpreter will perform complex tasks like data analysis, visualization, math and coding. Bing search is also getting upgraded to GPT-4 for better results on more complex topics.

Google Debuts Gemini AI in a Video Featuring “Generative” Editing

Google debuted its answer to Microsoft-backed OpenAI’s GPT-4, the new Gemini generative AI model. Gemini is natively multi-modal model, so it can analyze text, audio, video, images, and code and produce better results for handwritten content, images, and videos. simultaneously For example, Google showed a video of Gemini reading a student’s hand-written math homework, and explaining where the student missed the question and why. Instead of a live demo, Google released a video showing Gemini interacting with a person and responding to verbal questions. But behind the scenes Microsoft actually fed the AI text prompts and then edited the video for effect. While some oppose laws limiting AI, evil Microsoft is reason enough it’s needed.

Even After Deaths and a Recall There Are No Safeguards Against Using Tesla Autopilot on Unsuitable Roads

Tesla on autopilot

Tesla’s Autosteer feature is only intended for use on controlled-access highways with a center divider, clear lane markings, and no cross traffic. Tesla also says not to use autopilot on hilly roads or those with sharp curves. But after nearly a thousand wrecks involving Autopilot Tesla is recalling all 2 million cars with the feature. The “over the air update” will include additional alerts and monitoring “to encourage the driver to adhere to their continuous driving responsibility whenever Autosteer is engaged.”

But there is still nothing stopping drivers from using it on rural roads and city streets where it is not intended. Tesla says it’s not liable for accidents because it is up to the driver to determine when it’s safe to use the feature, even though the car’s GPS knows exactly what kind of road it’s traveling.

The EU Passed Laws Regulating AI

EU judges regulate AI

The EU’s AI Act is touted as the world’s first comprehensive AI law, aiming to regulate artificial intelligence to foster innovation while ensuring safety, transparency, and fairness. The Act classifies AI systems by risk, imposing obligations varying with risk levels, and sets specific provisions for high-risk and general-purpose AI, including transparency and evaluation requirements.