Projects - Ideas - And Other Insanity
A method and application to analyze real world Risk and Opportunity.
One of the biggest hurdles of astrophotography is processing the images you obtained from your telescope and camera. Take a few moments and see how AI can speed up your processing taking only 5 minutes or so.
This proposal explores whether artificial intelligence could be used as a TRAINING AID for police investigative interviewing, rather than as a tool for determining guilt, credibility or deception. Using synthetic case files and volunteer role-players, trainees would conduct simulated investigative interviews while an AI system assists with transcription, chronology, comparison with known facts, identification of contradictions, unanswered questions and possible follow-up opportunities. After the interview, instructors could compare the AI analysis with their own assessment and determine whether the technology adds meaningful value to the training process. The objective is simple: **does AI-assisted analysis help investigators recognize important inconsistencies and develop better follow-up questions?** The system would remain entirely subordinate to the instructor and investigator. No real suspect, active investigation or confidential police information would be required during the initial proof of concept. If the technology fails to improve training performance, the experiment ends. If measurable benefits are demonstrated, police services could then decide whether further research into more advanced applications is justified.
This proposal explores the use of artificial intelligence to assist journalists during live political interviews and press scrums. A specialized camera and microphone system would capture high-quality video and audio while AI processes the interview in real time. The system would immediately transcribe statements, check factual claims against reliable sources, compare answers with previous statements, identify unanswered questions or changes in wording, and note significant departures from the subject’s established communication patterns. Information would be presented discreetly to the reporter through smart glasses or an earpiece using a simple system: **Green — Supported or adequately answered** **Amber — Requires clarification or verification** **Red — Significant contradiction or unresolved issue** The AI could then suggest concise follow-up questions, allowing the reporter to challenge unsupported claims while remaining focused on the interview. In a press scrum, verified information and unresolved questions could also be shared among participating reporters. If one journalist receives an evasive answer, the issue would remain available for another reporter to pursue. The technology would not be presented as a lie detector. Facial expressions, speech changes, pauses, gaze, and other observable behaviour would provide only secondary information. The strongest conclusions would continue to come from factual evidence, contradictions, and the responsiveness of the answer. The objective is straightforward: give journalists better real-time access to facts, institutional memory, and effective follow-up questions so important issues cannot be easily avoided simply by changing the subject or moving to the next reporter.
A Safety Initiative seeking to improve Pedestrian Safety within intersections absent of traffic signals. Enhanced methods of traffic enforcement directed toward problem intersections having abnormally high vehicle/pedestrian collision rates might reduce this occurrence.
We often hear drivers proclaim that automated enforcement is less successful than manned enforcement. This is somewhat misleading, since continuous monitoring of an intersection will naturally generate far more tickets than could be issued through manned enforcement alone. Automated enforcement has also been shown to significantly reduce collisions. That said, many problem intersections may not warrant the cost of a permanent monitoring system. Selective traffic enforcement is intended to alter the behaviour responsible for collisions. Where a relatively small number of repeat drivers account for a disproportionate share of dangerous conduct, identifying those drivers and applying conventional demerit-point consequences may produce a lasting deterrent after the enforcement unit has moved elsewhere. Permanent and prominent signage advising drivers that an intersection MAY be monitored could further increase awareness and provide a continuing deterrent. Active enforcement, however, should not be advertised. This is not automated enforcement; it is enhanced manned enforcement. Police officers do not ordinarily announce their presence when temporarily monitoring a problem location. Because the portable system would have a relatively small physical footprint, it could be deployed discreetly and moved easily between locations. This should increase the likelihood that offending drivers are observed under normal driving conditions, allowing enforcement to be directed toward the behaviour creating the risk. The success of such a program must be judged by reductions in dangerous driving and collisions, rather than enforcement volume or revenue.