Chat-GPT 4’s New Voice

In recent months, Chat GPT has been evolving. It has gotten a new voice mode with a surprisingly life-like sound too it. The program can take and respond to real time requests. It is even able use the camera to ‘see’ the user and their surroundings. This is a massive jump in the programs’ complexity and a large step forward for AI interactivity. Open AI is hoping to make this a gamechanger for personal AI assistants

Unlike the expected monotone and robotic voice AI usually makes us think of, this new feature adds in emotion and tone to the voice. It appears to be able to respond in real time, and can laugh and do impromptu impressions or even sing. Open AI is doubling down on making Chat-GPT more than just a chatbot. The ability to reason over voice and video plus being able to react to interruptions and new information, is part of their plan to make it far more like a high powered virtual assistant.

Open AI has been striving towards this goal for some years now. Adding more and more functionality. Now they seem to be on the cusp of their next milestone. AI you can talk to and have a conversation with, like something right out of science fiction. The general public wont be getting this feature yet thought, there is still a lot that needs to be worked out.

Namely all the potential for misuse of the newly evolved program. They are working hard to make sure it can’t be used to falsely create problematic content or violate copyright. So only certain chosen testers are allowed access to it. As with any new feature there are also likely tons of hiccups and bugs to work out too. But From the videos Open AI has put out, it could be a massive shift in human AI interaction.

AI Powered Health Monitoring

Artificial Intelligence advancement is changing many areas of our lives. After completely upending the IT Sector it has moved its way over to the health industry. Many smartwatches or other fitness monitors have been turning to AI to interpret all the data they gather. turning what is normally simple devices for step count and heart rate into something more.

Technicians seek to allow these devices to not only take in the raw info from the sensors, but also process it with AI. Meaning the device could be used to give you a snapshot of your health throughout much of your body. Not only that, it will also be used to predict or warn about health concerns before they happen. With the data you could go to your doctor before something bad happens. Once there your wearable could give them the full rundown on what is happening inside your body. Or even alert your physician if it a serious condition, though that brings up privacy concerns.

For example, with the data they collect, smart watches can predict cardiac issue with around a 97 percent accuracy. AI powered devices have also shown success in improving sleep, and fitness routines. With their ability to track your health data and activity it can guide your fitness plans from an outside perspective. These apps could tell you what your missing, if you need to eat more, if you over did it, or if your at risk of something.

There are still challenges, like privacy of medical data, or security. Also the battery and equipment is on the edge of what is possible on such small platforms. Though once the engineers get to solving those issues, these devices will likely be everywhere. AI wearables will carry loads of info about what’s going on inside our bodies. They will absolutely be incredible tools to help us be healthier, but the AI program that reads and reports the data will know tons about you. That is something companies will likely have to reassure customers about.

The Pentagon’s Mission to Hunt Vulnerabilities

DARPA, the Defense Advance Research Projects Agency. These are the Pentagons special development group. They have been backing tons of cutting edge research of the years. Anything you have ever seen in a Sci-Fi movie they are probably trying to make it. Nowadays they are firmly focused on AI. Recently they have challenged over 90 teams to work on and build a AI based vulnerability detector.

Not as glamorous sounding as it is amazing, but this tool could be groundbreaking. Currently the challenge is to probe all open source codes and find issues then fix them automatically. Many open source codes are massive projects and used by loads of people around the world, things like Linux. In fact during this competition the team found numerous vulnerabilities in Linux. Each vulnerability in these codes could cause massive damage to individuals or companies. As the code gets larger and more complex it gets much harder to find all the holes to patch up.

That is where DARPA seeks to step in with their Cyber challenge. They want a team to develop a program that will detect and patch all of these vulnerabilities. With their backing they seek to support the most promising programs which can benefit the public. The Pentagon is focused heavily on cybersecurity these days and this program could lead to the development of a new weapon in the governments fight against hackers.

It seems like DARPA is happy with the progress they have made so far. They were not sure if it was possible and the fact they they succeeded in finding vulnerabilities is the proof they wanted. They awarded around 30 million dollars to the top teams and want they back next year. Should this type of system really get into the public space it could make a lot of peoples data much more secure. Identifying Zero day vulnerabilities before anyone can use them. Once AI can be trained to do this it will also probably change Cybersecurity forever, hopefully making it much easier!