Skip to main content

From Narrow AI To Smart Cities – the overreach of the Tech Sector

 



From Narrow AI Tools – to the design, development, deployment, and management of industrial metaverse applications

We can look at AI applications as tools. This point of view though is far too narrow. These tools are unlike anything we have utilised so far. NVIDIA are using the term ‘Omniverse Cloud’ to name it’s platfom-as-a-service offering , that provides ‘ a full-stack cloud environment to design, develop, deploy, and manage industrial metaverse applications.’

To put more simply: it’s like a virtual workshop where people can design, create, and manage things like manufacturing robots, buildings, and even whole cities (good luck with that last one).

Whilst I can readily envisage how PAAS system can function efficiently in the marketing sector, and to a large extent, manufacturing sectors and perhaps even buildings, I fail to see the possibility of it extending much further. Barry Smith in his talk about urban planning and smart cities provided a strong critique of the issues involved.




The multitude of risks involved in any wide deployment of AI on a city-wide scale would seem insurmountable. There are many papers that highlight the dreams (rather than aims) of Smart City projects and the challenges they may present. I’ll just quote from one: ‘Smart Cities: Opportunities, Challenges, and Security Threats’, Khalifa.

The introduction sets out the dream:
‘Smart cities represent a lifestyle totally based on making use of such unprecedented technological developments as Artificial Intelligence systems, the internet of things and big data, with the aim of maximizing the use of the available resources, reducing energy consumption and waste, creating an environment that enhances creation and innovation, and improving the quality of life for people by reducing the cost of living and making life easier and safer.’

At the end of the paper Khalifa sets out the Traditional Threats, which he expands upon under the titles:

  • Targeting Critical Infrastructure and Government Assets
  • The Threat to Banking and Finance Systems
  • Hacking the Internet of Things
  • Targeting Artificial Intelligence Systems
  • Hacking Cloud Storage Platforms
  • The Emergence of New Types of Crimes
  • Technical Problems in Software and Devices
There is then a look at Non-Traditional Threats. This is split into three: ‘The Possibility of Societal Resistance to Change’, ‘Violating People’s Privacy’, and ‘Jeopardizing Democracy’.

Unfortunately I consider that in two of these areas the arguments are not as pointed or as insightful as they could be. So I’ll just quote one of them:

Violating People’s Privacy

'The privacy of individuals is one of the most controversial issues concerning smart cities. The citizens' data, digitized and stored on smart phones, clouds, etc., are always in danger of violation from inside the city or from outside it, either by organized crime groups or by other countries. Credit card information, GPS information, biometric data, medical data, etc. are always available for the companies that operate the smart city. Besides, people may feel uncomfortable because of the security cameras that would meet them wherever they go, hence the classical question; must we jeopardize people's security to protect their freedom or must we restrict their freedom to ensure their safety? (Elmaghraby, 2014).'

Whole papers have been written and devoted to privacy concerns of tech, so Khalifa was never going to be able to encompass much of the literature in the space devoted to this one issue in a paper. Even so it is sufficient information to ask questions about. 

All 'Smart' systems, be they your smartphone, smart TV, smart home, or indeed smart workplace or smart city of the future rely upon data. The more the better. It's this big data, and AI's increasing capabilities of analysing and connecting disparate sources together that should be of concern. The EU data act and GDPR are, supposedly, a barrier to the miss usage of data, with the 'right to erasure' - or the right to be forgotten as it's often represented - appearing central to these. 

To see how effective, or ineffective, the regulators response to GDPR breaches have been you can examine the ICO website. 

The last action, at the time of writing, was a 'reprimand' by the ICO of Thames Valley Police (30th May, 2023:

'This investigation has found that TVP have inappropriately disclosed contextual information that led to suspected criminals learning the address of a witness (the data subject). As a result of this incident, the data subject has moved address and the impact and risk to the data subject remains high.

This incident occurred because TVP did not have the appropriate organisational measures in place to ensure that their officers were aware of existing guidance around disclosure and redactions. TVP have been unable to evidence that the officer who responded to the information request from the housing authority had received redaction training or was aware of existing policies around sharing information.

Further to this, there was no oversight of the redaction process as TVP thought that the officer in question had sufficient experience to complete the redactions.'

The slap on the wrist, the result of the reprimand, is minimal, to say the least. No justice, no peace comes to mind.

Risk, privacy and such matters can seem very abstract concepts, until, as in this case, a trusted authority mishandles information to jeopardise life and turn your world upside down. 


Labels:

Comments

Post a Comment

Popular posts from this blog

The AI Dilemma and "Gollem-Class" AIs

From the Center for Humane Technology Tristan Harris and Aza Raskin discuss how existing A.I. capabilities already pose catastrophic risks to a functional society, how A.I. companies are caught in a race to deploy as quickly as possible without adequate safety measures, and what it would mean to upgrade our institutions to a post-A.I. world. This presentation is from a private gathering in San Francisco on March 9th with leading technologists and decision-makers with the ability to influence the future of large-language model A.I.s. This presentation was given before the launch of GPT-4. One of the more astute critics of the tech industry, Tristan Harris, who has recently given stark evidence to Congress. It is worth watching both of these videos, as the Congress address gives a context of PR industry and it's regular abuses. "If we understand the mechanisms and motives of the group mind, it is now possible to control and regiment the masses according to our will without their
Post a Comment
Read more

Beware the Orca, the challenge to ChatGPT and Palm2 is here

  So Google's 'we have no moat' paper was correct. If you train an LLM wisely then it's cost effective and cheap to produce a small LLM that is able to compete or even beat established, costly LLMs, as Microsoft has just found. It's another excellent video from AI Explained, who goes through some of the training procedures, which I won't get into here. Orca, is a model that learns from large foundation models (LFMs) like GPT-4 and ChatGPT by imitating their reasoning process. Orca uses rich signals such as explanations and complex instructions to improve its performance on various tasks. Orca outperforms other instruction-tuned models and achieves similar results to ChatGPT on zero-shot reasoning benchmarks and professional and academic exams. The paper suggests that learning from explanations is a promising way to enhance model skills. Smaller models are often overestimated in their abilities compared to LFMs, and need more rigorous evaluation methods. Explana
Post a Comment
Read more

What is happening inside of the black box?

  Neel Nanda is involved in Mechanistic Interpretability research at DeepMind, formerly of AnthropicAI, what's fascinating about the research conducted by Nanda is he gets to peer into the Black Box to figure out how different types of AI models work. Anyone concerned with AI should understand how important this is. In this video Nanda discusses some of his findings, including 'induction heads', which turn out to have some vital properties.  Induction heads are a type of attention head that allows a language model to learn long-range dependencies in text. They do this by using a simple algorithm to complete token sequences like [A][B] ... [A] -> [B]. For example, if a model is given the sequence "The cat sat on the mat," it can use induction heads to predict that the word "mat" will be followed by the word "the". Induction heads were first discovered in 2022 by a team of researchers at OpenAI. They found that induction heads were present in
Post a Comment
Read more