02.11.2018 When and Where will a Breakthrough Come? Technology - - PDF document

02.11.2018 1

INF3490/INF4490 Biologically inspired computing

Future Perspectives on Artificial Intelligence – What to expect and should we worry? Jim Tørresen, October 31, 2018

When and Where will a Breakthrough Come?

• Technology breakthroughs often happens

randomly and not linked to major initiatives and projects.

• AI breakthroughs depend on the invention of

scalable learning methods.

– Need to understand more about how scalability and complexity arises in nature.

• Contributors to progress in AI:

– Researchers mimicking biological or medical phenomena. – Researchers solving engineering problems

2

CAM (Cellular Automata Machine) Brain

3

What Methods are Best?

4

Autonomous Systems

• Systems that can operate with no or limited

human control

• Degrees of human control:

– Full autonomy (human decides only the goal) – Remotely monitored (autonomous planning but human take over if the system is misbehaving) – Remotely operated (autonomous steering using human planning) – Remotely controlled (human take all decisions)

5

How Similar to Humans?

• Similarity depends on progress in a number
• f fields such as AI methods, computing

power, vision systems, speech recognition, speech synthesis, human-computer interaction, mechanics and actuators.

• Design and usability will be essential for

future robots but they should not necessarily be as similar as possible to humans (ref. uncanny valley).

• We will see robots having human-like

behavior, but not human consciousness.

6

02.11.2018 2

In what ways can AI be helpful/good and what are the potential challenges?

Input collected from students attending the class:

• helpful/good:

– Eliminate repetitive tasks for human – Solve diff problems – Better NLP

• potential challenges:

– Legal consequences – Personal data sharing – Eliminating jobs – Aligning goals with humans

7

Is Terminator Coming Close?

8 9

“Humans, limited by slow biological evolution, couldn’t compete and would be superseded by A.I.” AI is our “biggest existential threat” I am in the camp that is concerned about super intelligence.

Ethical Risks of Developing AI Systems

• People may become unemployed because of

automation.?

• We get too much free time.?
• Artificial intelligence can be used for

destructive and unwanted tasks.? (ref. “1984”

George Orwell)

• Successful KI can lead to the extinction of

mankind?

10

Future Scenario with Autonomous Interacting AI Systems

11

Ethics for Artificial Intelligence

12

• Designers must be aware of possible ethical

challenges that should be considered

– eg. protect sensor data, avoid misuse, and allow human inspection of functionality

• The systems should themselves be able to

do ethical decision making to reduce the risk

• f unwanted behaviour
• Optional paper:

https://www.frontiersin.org/articles/10.3389/frobt.2017.00075/full

02.11.2018 3

Ethics for Programmers

• In the book “Moral Machines” it is argued that ethic

competence should be included during program development by e.g.

– machine learning methods based on examples of ethical and unethical behavior

• Software that will replace human evaluation and

social function should adhere to criteria such as

– accountability (no. ansvarlighet) – inspectability – manipulation robustness – predictability (no. forutsigbarhet)

13

Self learning/Machine learning

System to be designed Data set/ specification Algorithm Glass box Black box vs

Somewhat transparent and understandable behaviour

What would you require of a robot to be used in your home?

15 16

Ethical Guidelines for Robots and their Developers (Asimov 1942)

• 1. A robot may not harm a human being,
• r through inaction, allow a human to be

injured.

• 2. A robot must obey orders given by

human beings except where such

• rders would conflict with the first law.
• 3. A robot must protect its own

existence as long as such protection does not conflict with the first or second law.

17

Ethical Recommendations for Commercial Robots (Euronet Roboethics Atelier)

• Safety: There must be mechanisms (or opportunities

for an operator) to control and limit a robot's autonomy.

• Security: There must be a password or other keys to

avoid inappropriate and illegal use of a robot.

• Traceability: Similarly as aircraft, robots should

have a "black box" to record and document their own behavior.

• Identifiability: Robots should have serial numbers

and registration number similar cars.

• Privacy policy: Software and hardware should be

used to encrypt and password protect sensitive data that the robot needs to save.

18

02.11.2018 4

This lecture is based on a translation of the last chapter in an introductory book to artificial intelligence

19

Innhold:

• Kunstig intelligens og

intelligente systemer

• Problemløsning med

kunstig intelligens

• Evolusjon, utvikling og

læring

• Sansing og oppfatning
• Bevegelse og robotikk
• Hvor intelligente kan og

bør maskiner bli?