AI models….are we all doomed?

I dont think AI models on their own are dangerous to humankind as claimed by many. They only become potentially harmful when we give models access to the tools to take action…i.e. APIs

So I asked ChatGPT to write me an article on this to save me time!

Title: The Risks of Unrestricted API Access for AI Models: Safeguarding Humanity's Future

Introduction:

Artificial Intelligence (AI) has witnessed remarkable progress, transforming industries and revolutionizing the way we live. However, the uncontrolled access granted to AI models through APIs (Application Programming Interfaces) poses significant risks. This article delves into the potential dangers associated with AI models having unrestricted API access and highlights the importance of establishing robust safeguards to protect humanity's future.

The Power of AI Models:

AI models are designed to analyze vast amounts of data, uncover patterns, and generate valuable insights. They possess immense potential to optimize processes, enhance decision-making, and offer unparalleled convenience. However, it is crucial to acknowledge that AI models, no matter how advanced, lack inherent ethical judgment and the ability to foresee the consequences of their insights.

The Role of APIs:

APIs act as gateways that enable interaction and data exchange between AI models and external systems or services. They provide a means for AI models to access and utilize various functionalities, granting them a broader scope of action. While APIs facilitate seamless integration and efficiency, they also introduce risks when coupled with powerful AI models.

Unintended Consequences:

When AI models are given unrestricted access to APIs, the potential for unintended consequences arises. The models may generate insights that, when acted upon, result in negative outcomes or harm individuals or society. This risk is amplified when AI models are allowed access to critical systems such as financial trading platforms, autonomous vehicles, or even weapon systems.

Ethical Concerns:

Granting AI models unrestricted API access raises ethical questions. Who bears responsibility when AI-generated insights lead to adverse effects? Can we trust AI models to act ethically and in the best interest of humanity? The complex interplay between AI, APIs, and human decision-making demands careful consideration and proactive measures.

Mitigating the Risks:

To mitigate the risks associated with AI models and API access, several measures should be implemented:

1. Robust API Governance: Implement strict controls, authentication mechanisms, and access restrictions to ensure that AI models can only interact with APIs that align with ethical standards and regulatory frameworks.

2. Ethical Guidelines: Establish clear ethical guidelines and standards for AI model development, deployment, and usage. These guidelines should address the limitations of API access and the responsibility of human decision-makers when acting upon AI-generated insights.

3. Regular Auditing and Monitoring: Continuously monitor AI model behavior, API usage, and the impact of generated insights. Regular audits can help identify potential risks and intervene when necessary.

4. Human Oversight: Retain human oversight and decision-making authority when implementing AI-driven systems. Humans must have the final say in acting upon AI-generated insights, ensuring that ethical considerations and human well-being are prioritized.

5. Collaborative Efforts: Encourage collaboration between AI researchers, developers, policymakers, and ethicists to collectively address the challenges posed by unrestricted API access. A multidisciplinary approach can foster comprehensive solutions that balance technological advancement with societal well-being.

Conclusion:

While AI models offer immense potential, it is crucial to recognize the risks associated with granting them unrestricted API access. Ensuring AI models operate within ethical boundaries and align with human values is paramount. By implementing robust governance frameworks, ethical guidelines, and human oversight, we can harness the power of AI while safeguarding humanity's future. As we navigate the evolving landscape of AI and APIs, responsible decision-making and proactive measures are essential to mitigate risks and harness AI's transformative potential for the greater good.

Identity a solution?

You may remember my post from May 2018 where I talked about identity being the future of our progress in the digital world.

Maybe there is an answer….I was listening to a TWIT podcast a few days ago and I heard a group of technology security experts talking about exactly what I was discussing in my post.

One of the panelists was Steve Gibson who has been a long time host of the Security Now podcast also on the TWIT network

Steve has been working on an identity solution called SQRL (Secure Quick Reliable Login) which, in very simple terms, put the onus on client sites to obtain identity from the user via unique public/private keys generated for each domain name. Having read the specification I really this, or something like it, could be the answer to our problems. The fundamental idea is the the user provides identity details, it is not stored in a central database owned by someone we are unlikely to ever fully trust.

Take a look….and see what you think

Securing Identity is the key to our future

I was riding home on a packed commuter train the other day and decided I would conduct an experiment. What could I find out about the personal details of one of the people in my seating area of six just by using my phone and whilst on the train. One of them happened to be reading a printed email with their name on so I went to work.

Within a couple of minutes I had worked out which town they lived, which company they worked for, what they did for that company. I also knew their interests, hobbies and saw photos of the persons last summer vacation. I also knew their partners and kids names and where some of their relatives lived. It was simple because I could work out that the person most likely lived in New York or Connecticut given the train we were on. That, with their name is all I needed.

Now I like to think I am an honest person but given more time and a more nefarious intent I am sure I could have got into a lot more detail on this person. For example where exactly they lived, how much their house cost, how much debt they took out for the house etc etc....

This small story gives a clear picture of power and the curse of the age we live in. We have access to data unlike any time in human history. Having been fortunate to have been a teenager at the very beginning of personal computing it is so exciting to me see my dreams in college come to fruition. Online unlimited information - encyclopedias, maps, videos, communication, collaboration it is quite amazing what has happened in the past 30 years.

This explosion in power can be used to do so much good as well as bad. As human beings we will surely go down all avenues of opportunity.

At the root of our future in my view is the sanctity and value of our identity. I am afraid with all the great things technology has achieved in the past few decades we have not solved how to be sure of and protect someone’s identity. 

What do I mean by identity? Some questions for you to understand....
How many passwords for web sites do you have to remember?
For those important sites how often do you have to deal with the hassle of 2 factor authentication? 
When you want to buy something online how often do you have to enter your credit card online?
How many times has your credit card been replaced because your card company tells you there has been a fraud? Has the number of times this has happened increased in the last couple of years?
Have you figured out yet that anything you get for free on internet is actually just selling data about you so they can advertise back to you?
Has it ever occurred to you that in times of pervasive technology you still have a paper/plastic passport and driving license and you have to queue up, often for an extended time, to have it ‘checked’ by a human being?

With all of these issues we desperately need a simple, near foolproof, flexible way to secure and, when you want, provide items that make up our identity - name, date of birth, address, links to our finances, nationality, etc etc

I am sorry, dear reader, I dont have the answer to this problem but I fear we wont get a resolution soon because it is not in the interests of the current slew of technology companies for us to have this important feature.

Should such a mechanism be run by the government...which one? 
If a company were to run this would we be willing to pay for it? VERY UNLIKELY
If we got it for free how would it be paid for? ADVERTS PROBABLY

If you have any ideas let me know!